the impact of the common agricultural policy on
TRANSCRIPT
University of RichmondUR Scholarship Repository
Honors Theses Student Research
2008
The impact of the common agricultural policy onagricultural productivityFlemming Schneider Rhode
Follow this and additional works at: http://scholarship.richmond.edu/honors-theses
This Thesis is brought to you for free and open access by the Student Research at UR Scholarship Repository. It has been accepted for inclusion inHonors Theses by an authorized administrator of UR Scholarship Repository. For more information, please [email protected].
Recommended CitationRhode, Flemming Schneider, "The impact of the common agricultural policy on agricultural productivity" (2008). Honors Theses.Paper 369.
The impact of the Common Agricultural Policy on Agricultural Productivity
By
Flemming Schneider Rhode
Honors Thesis
in
Department of Economics
University of Richmond
Richmond, VA
April 281h, 2008
Advisor: Dr. Gabriel Asaftei
Abstract
In 2004 10 new countries, primarily Eastern European countries, joined the European Union(EU)
with much media speculation concerning how this would affect their economies. This paper will
look at how the Common Agricultural Program( CAP) affects agricultural productivity by
measuring four independent variables and by using the 2004 entry as a natural experiment. The
paper will estimate how the CAP affects the average farm size, fallow land area, proportion of
farming dedicated to organic farming, and GOP growth. This is expected to impact agricultural
productivity through increasing returns to scale, input availability, efficiency ofland use and the
Environmental Kuznets Curve respectively. The findings indicate an overall increase in
agricultural productivity as a result of joining the EU from higher average farm size, lower levels
of fallow land and higher proportions of organic farming. The estimation technique used was an
ordinary least squares regression with fixed effects for each country.
I - Introduction
This project will look at the impact of subsidies on agricultural productivity. Specifically
my analysis will look at the CAP in the EU and how it affects the productivity of the agricultural
sector EU15 (the original EU member states) and EU8 (The 8 Central and Eastern European
countries that joined in 2004). Specifically the variables examined will be GDP per capita (as the
countries converge with the "old" EU countries' GOP level, pollution and pesticide levels may
change and thus affect productivity), land use (how will mandatory fallow land policies affect
productivity), proportion of organic farming (how will agricultural subsidies incentivizing
organic farming affect productivity), and increasing returns to scale (how will the Euro per
hectare policy affect the average farm size).
Investigating the impact of subsidies on agricultural productivity in Central and Eastern
Europe provides a very interesting case study of agricultural subsidies. One of the reasons for
this is that all the Central and Eastern European countries joined simultaneously in 2004 and
received economic assistance via the CAP from day one. This natural experiment presents a
unique opportunity to investigate whether or not the CAP has had a negative or positive impact
on agricultural productivity. This is also a very important issue as the CAP is engulfed in
controversy and contention. This is both from within the EU as net contributors demand a
reduction in the program as well as from international organizations such as the World Trade
Organization charging that the CAP is an instance of unfair protectionism. This topic is also
worth investigating because ofthe sheer financial magnitude of the pro!:,'Tam, in 2005 a total of€
43 bn was spent maintaining the program (BBC News 2005). Findings indicating an increased
level of productivity due to inclusion in the EU and access to the CAP would help justify the
program's continued existence. Conversely, findings indicating that the CAP lowers productivity
would lend credence to the critics of the program advocating reducing the subsidies or
restructuring the program. Finally, the CAP itselfbecame one of the main problems in the
negotiation process of accepting the new members of the EU because they did not receive the
full etiro per hectare support from day one. The CAP is thus a very significant policy possibly
affecting the future of EU expansion.
The CAP itself is also an interesting example of an agricultural subsidy as it has different
characteristics, some of which would be expected to reduce productivity while others would be
expected to increase productivity. This will be discussed in further detail in the theory section.
Brief History of the Common Agricultural Program in the EU
The EU (previously known as the European Coal and Steel Union) formed a uniform agricultural
program for all member states in the late 1950's with the objective of providing farmers with a
decent income, consumers with quality products at low prices, and preserving the rural heritage
of Europe. The CAP later adopted has become a key component of the pan-European agricultural
policy, and was destined to secure food supply for the European Continent. This program was
seen as critical in the aftermath of the Second World War when food security was still a concern.
While the main part of the CAP was a subsidy scheme to support agricultural prices, an
important component of the program was also incentives to encourage free trade among the
member countries. In time, the CAP became the main part of the EU budget (untill992 it took
up 62% of the EU budget). The regional (national) differences between net beneficiaries and net
payers of this program made the CAP quite contentious. As a result there has been both internal
as we11 as external pressure (e.g. the World Trade Organization) to change the policy. However,
since a relatively large majority of votes (and in some cases outright consensus) is needed,
structurally changing the CAP has proven time consuming and challenging.
The initial process of negotiating a CAP was difficult, it was first proposed at the Treaty
of Rome in 1957, but not ratified until 1962. The problematic part was to have a common market
while at the same time offering the same level of protection to the farmers which was enjoyed
under the national level subsidy programs. Specifically, the compromise was reported to be
between Gennany and France; Germany's industry gained access to France, but Germany was to
help finance the subsidies paid to French fanners. The accuracy of these reports is hard to verify
as the negotiations were behind closed doors. It is, however, beyond any doubt that Germany
was and still is the main contributor to the CAP and France is the biggest beneficiary with its
large agricultural sector. Today with the accession of the 12 new countries, France will for the
first time in history become a net contributor.
There have been several stages of development within the CAP. The year 1992 marked a
breakthrough year as the MacSharry reforms were introduced. The main impact of these reforms
was to bring prices closer to equilibrium levels and minimize overproduction. Instead of paying
for overproduction of goods which led to the infamous sugar mountains and wine lakes which
. ' were expensive to maintain, the MacSharry reforms paid to create reforestation, more fallow
land and direct payments to farmers to ensure retirement. One of the main forces behind this
change was the World Trade Organization's Uruguay Negotiation Round forcing EU countries to
have a more reasonable agricultural policy which would not adversely affect agricultural sectors
.in other countries. Previous EU policies of floor prices created a glut of food products which was
dumped on primarily developing countries labeled as "emergency food relief' but ended up
destroying farmers' livelihood.
Between May 2004 and January 2007 the EU included 12 new member states, most of
which come from Eastern Europe and are currently in transition economics. This was a very
controversial move in many different circles. There was much skepticism within "old" EU
member states as there was a fear it would be expensive to subsidize the new member states'
agricultural sectors. The media also paid close attention to the economic impact of this
enlargement. The negotiated consensus was a step-wise integration of the Eastern European
countries in the CAP. However, in the negotiations it became clear that large Eastern European
agncultural countries (in particular Poland) were unsatisfied with the slow full implementation of
the CAP toward the new member states. After long negotiations the I 0 Eastern European
countries have been admitted into the.EU and correspondingly as beneficiaries of CAP.
The next section of this paper will outline the research question and some ofthe literature
and theoretical work already done on this topic, followed by a discussion of the theoretical
background, the model used, and the data used for this assignment before results are presented
and some concluding remarks arc made about the findings.
II - Literature Review
A large number of both books and published articles have examined the issue of agricultural
productivity in a variety of subtopics such as geographic locations or specific commodities
(listed below). Even though most of these articles focus on a very specific subsidy within a given
economy there arc some universal characteristics. There is also a consistent subset of literature
concerning the EU expansion of the CAP itself and how this affects the level of productivity
within the old as well as new economics in general and whether there is a convergence of growth
rates. A third set of literature is specific to the use of pesticides and how this affects both short
run and long run productivity.
Subsidies in general arc thought to create inefficiencies in economic systems as the price
signals they send can distort markets and their optimum conditions. With regards to agricultural
subsidies in particular Hu and Antle (1993) find that there is an optimum level of productivity
which can be significantly affected by agricultural subsidies. In moderate levels government
policies can positively affect agricultural productivity due to innovations in crop yields,
specialization and economies of scale. However, very high levels of subsidies have a negative
atTect due to the inherent high tax rates, resource and technology constraints as well as the
massive distortions to the incentives of farmers due to irrational resource allocation. This study
was based on global World Bank data collected in the 1960's to the 1980's. Agricultural
subsidies also incentivizes a certain type of agricultural structures, Breustedt and Glauben (2007)
examine how higher levels of subsidies results in lower levels of exits, thereby indicating that
stagnant high levels of subsidies would slow down the naturally changing composition of
agricultural sectors, while it maintains that this is the case to a lesser extent with smaller farms.
Relating specifically to the EU, the 2004 and 2007 enlargements have long been expected and a
wide range of predictions have been made on the effect on the EU economies. Baldwin ct a!.
(1997) assess the impact as exclusively positive for both entering countries as well as old
countries from bigger markets, although entering countries would benefit the most. Pctrakos ct al
(2005) look at the distribution ofbenefits among the countries, and conclude that there will be a
core/periphery gap where the countries geographically closer to the West will benefit
disproportionately. Leguen de Ia Croix (2004) in an official paper of the European Commission
indicates the productivity of the entering Eastern European countries arc lower as they add 30%
of total hectare of agricultural land, but only add in between I 0-20% agricultural production,
though this does not control for input costs.
In terms of environmental aspects of the CAP, Van der Grijp and den Hond (2000) and
Serra et. al (2005) outline the significant reductions made in the I 990's of pesticides {kg/ha) used
by EU member countries due at least in part to tougher EU regulations. How this affects
agricultural productivity remains a contentious point. Grossman and Krueger (I 99 I) wrote the
seminalwork on the Environmental Kuznets Curve outlining how higher levels ofGDP would ,: f'
lead to lower levels of pollution. Managi (2006) applies this specifically to pesticides showing
tighter regulations abating the environment also removes decreasing returns to scale as many
farmers overuse pesticides and are thus forced to reduce their input costs and are forced closer to ':
the point of optimum use. Huang et al (2002) and Dasgupta et. al (2007) highlight cases in China
and Bangladesh respectively where lower levels of pesticides increased productivity as the price
of inputs were significantly reduced while the crop yields were not impacted significantly.
However, Shankar and Thirtle (2005) and Lansik and Silva (2004) determine that pesticides are
under-utilized in South Africa and the Netherlands respectively. The optimum use of pesticides
is very different according to the specific plant as well as climate, thus deviating findings do not
necessarily contradict.
Finally, another regulation of the CAP is that land needs to be fallow within a certain
timeframe and subsidies in the EU depend on the extent of diversification, Pascual (2003) finds
that diversification and leaving land fallow actually increases agricultural productivity in the
long run based off data in the Yucatan in Mexico.
This paper will take a look at these variables within the context of one specific
agricultural policy. The entry date is the same for all countries examined, so 2004 presents a
precise and uniform change in policy for EU8 while there should be no change in the agricultural
policies for the EU 15 countries. This paper thus presents a unique perspective to test the
applicability of findings of previous studies on EU member states.
III - Research Design and Theoretical Model
The CAP has a complex set of rules and regulations, but there are three components which each
have different impacts. The first of which is the grant to farmers in the EU for arable land in
good condition. In order to meet the condition of "good arable land" farmers must lay the land
fallow after a number of years dictated by the EU. Since this condition favours larger farmers
(Breustedt and Glauben 2007) due to increasing returns to scale, larger farmers can get larger
grants while keeping all other costs fixed. The increasing returns to scale could come from bulk
buying e.g. tractors and harvesters or from labour specialization on the farm. Thus the first type
of grant for having land in good arable condition is a function of its incentive to promote
increasing returns to scale. We would thus expect that inclusion into the EU would increase the
number of big farms and decrease the number of small farms because these larger farms, ceteris
paribus, have a competitive advantage. Because we would expect the average farm size to
increase we would simultaneously expect this to increase productivity and pivot the supply curve
to the right.
Figure 1 Prrce
Qlldl'ltit\
Moreover, farmers are mandated to, but also financially supported for, leaving part of
their land fallow every cropping season. The reasoning behind this is that in order to have good
soil the land should be left aside with no crops grown or animals grassing on it. This way the soil
will recover its nutrients and long term productivity is increased. However, because output from
this area is completely eliminated we would expect this in the short run to shift the supply curve
left and quantity supplied would decrease. At each and every price level producers would be able
to supply less output. The effect on productivity, in terms of income to farmers holding inputs
constant is somewhat more ambiguous. The reason for this is that elasticity of demand might be
inelastic. Food is generally considered a rather inelastic good as a whole, since there are no
substitutes. Of course American produced potatoes are substitutes to European potatoes, but
other protectionist components of the CAP (discussed later) effectively eliminate this as a viable
option for the consumer. Thus, if the demand indeed is inelastic then a leftwards shift of the
supply curve would reduce output, but increase income to farmers.
Figure 2
C)ual1 tit\ The second component of the CAP is the promotion of ethical farming as organic farming
is specifically rewarded with EU subsidies. Organic farming and ethical farming put further
restrictions on farmers as they are not allowed to use pesticides, synthetic fertilizers or plant
growth regulators. We would predict the CAP incentives for organic farming would increase the
proportion of organic farming. This in tum would reduce total output as a series of inputs are
now unavailable and the substitutes such as natural fertilizer are less efficient and/or more
expensive. However, the effects on productivity remain unclear. The reason for this is that while
the total output decreases, as in the number of potatoes harvested per hectare, the organic food is
typically more expensive and sold in a different market. Thus total output can go down and total
revenue and profits to the farmer go up because the price differential outweighs the reduced
output. However, in the absence of a subsidy program we would expect farmers to be profit
maximizing agents and grow either organic or non-organic crops on their land according to
which is more profitable. This would in the long run make each individual farmer indifferent to
the organic market and non-organic market because the price fluctuations in each commodity
would be equalized by the price adjustment mechanism. Thus, if the individual farmer is
indifferent to organic and non-organic farming in a world with no subsidies, a world in which the
CAP exists there is an extra incentive to do organic farming, there would be a shift the
proportion of farming dedicated organic farming away from the optimum point of allocative
efficiency. We would thus expect a leftwards rotation of the Supply curve due to higher costs of
production, even considering the higher price for organic foods.
Figure 3
Prict.>
01
As Managi (2006), Huang et. al (2002) and Dasgupta ct. al (2007) point out pesticides
are often overused 'so even if there is a reduction in total output, the larger reduction in costly
inputs actually increases productivity. If, however, pesticides are used close to its optimum point
or beneath it as Shankar and Thirtle (2005) and Lansik and Silva (2004) argue, further
restrictions on pesticides would lower productivity. As Central and Eastern European countries
join the EU a GOP convergence is·expected, while some predict this taking longer than others,
there is widespread consensus it would take place. According to the theory of the Environmental
Kuznets Curve, lower levels of GOP are associated with higher levels of pollution and chemical
use in agriculture, while higher levels of GOP are associated with lower levels of pollution and
pesticide use because of increased regulations as people value environmental protection higher.
The relationship, however, is not linear but quadratic as higher levels of GOP are associated with
higher levels of pollution at first, but then higher levels of GOP are associated with lower levels
of pollution. Depending on where the countries are on the Environmental Kuznets Curve,
increases in GOP could both cause higher and lower levels of pollution. As previously noted
some types of pollution restrict productivity while others do not so it is difficult to predict the
outcome of the direction of the rotation of the supply curve. It should also be noted that this
variable is most clearly identified in the long run, and thus the effects may not be visible in the
timeframe of this study.
Figure 4
Price
<)tt anti t\
In this model it is expected that farmers are profit maximizers and would seek to
maximize productivity under the constraints of the CAP regulations. This means farmers would
thus exploit certain regulations that would lead to less productivity, if it would lead to higher
profits via grants, but would, ceteris paribus, seek to maximize productivity. With regards to the
direct payments, this would function as an outwards shift of the supply curve as a higher amount
of money is rewarded for each good produced:
Price Figure 5
The reason why it is not certain the price would decrease is because the EU has
guaranteed to uphold a minimum price (PS) through the price maintenance policies which seck
to maintain a certain price level by buying excess production. While the influence of this subsidy
has been decreasing it is still in effect today. This functions as a minimum price creating a
surplus (PS-D- PC-S) which the government buys up. If the direct payment subsidies would
push prices down beneath the floor price, price would not decrease, regardless of how much the
supply curve shifts to the right. It should also be noted that this subsidy is only costly when in
fact the EU has to intervene and buy up excess surplus, and the price of the subsidy would be
amplified by the direct payments subsidy as farmers are not only paid to produce more, but the
extra production is purchased by the EU. Thus we would expect both the direct payments and the
price stabilization component of the CAP to increase quantity supplied, but there would also be a
corresponding increase in inputs used. Increased inputs used may be accompanied by economies
of scale from lower per unit costs from bulk buying, but it could also be the case that there would
be diseconomies of scale. Certainly there would be a dead weight loss to consumers as
graphically demonstrated, since the marginal cost curve is the supply curve and the marginal
benefit curve is the demand curve. The triangle DWL makes up this portion of the subsidy
measured amount of marginal cost excess of marginal benefit.
Figure 6
Pnce
PS
PE
PC
Qu(--u 1tit\
Some of the independent variables can be expected to counteract, we would expect larger
average farm sizes to increase productivity, but we would also anticipate a possible decrease in
productivity from increased proportions of organic fanning. These factors pivot the supply curve
right and left respectfully. With regards to quantity supplied we would expect the direct
payments and minimum price effects to increase quantity supplied, but the overall quantity
supplied might decrease if the supply curve pivots left enough to outweigh the rightwards shift in
the supply curve and possible disequilibrium effect. Because there is this ambiguity the test for
this study will be two tailed:
HO: The CAP does not impact agricultural productivity H 1: The CAP does impact agricultural productivity
The model used for this is a classical Cobb-Douglas production function:
(Eq.l)
Where Yu is the agricultural output produced by country i at time t, Ail is the Total Factor
Productivity which represents the residual impact on output not caused by labour or capital
inputs, L ail is the labour input, J(fl;1 is the capital input, and eu is the random unobserved error not
captured by the variables in the model. In order to run this equation in a regression and estimate
the exponents the equation has to be transformed into a Log-Linear form:
Ln(Y;J = In (A;J F (o. In (L;J fi In (K;J) + eu (Eq.2)
In this form the model can check for constant returns to scale and estimate the relative impacts
labour and capital have on output. More importantly, having isolated labour and capital we
would think of the different components of the CAP affecting Total Factor Productivity in the
following way:
Ln (A;J = Ao + Jc, In amount of fallow landit + A2ln proportion oforganic farmingit + Jc3 In average farm sizeu + Jc4 1n GDP per capitau + A5 EU variableit + eu (Eq.3)
We can then estimate the effect of these variables on output in the final equation:
Ln {Y;J = ).o + Jc, In amount of fallow landit + A2ln proportion of organic farming; 1 + Jc3 1n average farm sizeit + A4ln GDP per capitau + A5 EU variableu + o.ln Lu fJ In Ku +6; + eu (Eq.4)
The only differences in this equation are the addition oflabour and capital as independent
variables and the inclusion of 6; which represents the fixed effects.
IV- Data and Summary Statistics
The data available for this paper range from 2002 to 2005, the first two years of which the
Central and Eastern European countries were not member states of the EU. The data for
irrigation area, total farm labour force, number of holdings, and organic crop area had to be
interpolated. This was done by taking the arithmetic mean of the surrounding years to fill the
data gap, except for the labour force data, here the geometric mean was taken because intuitively
we would expect population and labour force growth to grow more rapidly than arithmetically.
There were also four countries eliminated from the sample: Estonia, Germany, Ireland and
Luxembourg which was due to excessive gaps in the dataset which could not be salvaged by
interpolation. This leaves the seven Central and Eastern European countries in the population
examined: the Czech Republic, Hungary, Latvia, Lithuania, Poland, Slovakia, and Slovenia, as
well as ll Western European states: Belgium, Denmark, Greece Finland, France, Holland, Italy,
Portugal, Spain, Sweden, and the United Kingdom.
The agricultural production data is measured in terms of millions of Euros at basic prices
which EuroStat defines as "the price received by the producer, after deduction of all taxes on
products but including all ~ubsidies on products. Output of the agricultural industry is made up of
the sum of the output of agricultural products, agricultural services and of the goods and services
produced in inseparable non-agricultural secondary activities." (EuroStat- Basic Price
Definition).
The amount of fallow land is measured in hectares (ha) and is the amount of arable land
in the crop rotation on which there is no crop growth and no animal grazing. The proportion of
organic farming is calculated by dividing the total area of organic crop production in ha by the
total area of crop production in ha. In order for crop area to be defined as organic the area has to
fulfill all criteria ofEEC regulation No 2092/91 which includes not using any synthetic fertilizer
or pesticide.
In order to find the average farm size the total crop area is divided by the total number of
holdings. The number of holdings is defined byEuroStat as "a single unit both technically and
economically, which has single management and which produces agricultural products. Other
supplementary (non-agricultural) products and services may also be provided by the holding"
(EuroStat- Holding Definition). The average farm size is thus measured in terms ofha.
' When controlling for the relative income of the country the data used in this study is
using the Gross Domestic Product per capita at market prices. It reflects the total value of all
goods and services produced less the value of goods and services used for intermediate
consumption in their production. It is also controlling for different price levels in different
countries by using the PPS (Purchasing Power Standards). The income is measured in Euros per
capita.
Irrigable area is chosen as a proxy for capital because it is closely correlated with other
proxies for capital such as tractors, harvesters and threshers. EuroStat defines irrigable area in ha
as "the maximum area which could be irrigated in the reference year using the equipment and the
quantity of water normally available on the holding" (EuroStat- Irrigable Area Definition).
The data available for labour is measured in terms of 1000 AWU (Annual Work Units),
where an A WU is the work of a full time employee. Finally the EU variable is a dummy variable
where a 0 value is given to non-member countries and a 1 to member-countries.
Table I presents some summary statistics outlining the arithmetic mean and standard
deviation for both the treatment group (EU8 -The newer members of the EU) and the control
group (EU15- the older members of the group). The most important things to note are that the
proportion of organic farming approximately doubled in the Central and Eastern European
countries after joining while in EU 15 it increased from 15% to 18%, that the proportion of fallow
land decreased dramatically upon entering for the EU8 members, and that the average farm size
increased marginally after the EU8 countries entered the EU, while it stayed static for the EU15
countries. Seeing a decrease in the amount of fallow land upon entering the EU and receiving
benefits for leaving land fallow seems curious. However, it must be noted in this context that all
countries prior to 2004 had their own national subsidies which could have had a higher incentive
to leave land fallow. Moreover, there is also the possibility that the general national subsidies
were not high enough to cover the variable cost of harvesting certain fields, and land was thus
laid fallow out of economic necessity.
Table 1- Summary Statistics
EUS EU15
2002-03 2004-05 2002-03 2004-05
Mean Standard Mean Standard Mean Standard Mean Standard Deviation Deviation Deviation Deviation
Agricultural 3713 4142 4230 4799 19320 19089 19281 19173 Production
(mio €)
Fallow 391 718 253 436 472 729 461 714 Land (ha)
Organic 5.88 7.96 10.82 8.54 15.21 6.87 18.82 12.82% Crop(% of
total)
Irrigable 89332 105551 76564 79889 1057399 1378236 1052457 1375446 Area (ha)
Labour 464 741 461 773 437 450 413 424 (1000 AWU)
Average 60 129 62 133 9 9 9 9 Farm Size
(ha)
GDP Per 12000 2874 13707 3849 23045 3279 24564 3535 Capita
The following graph depicts the agricultural production measured in mio. of Euros received by
all farmers in the countries divided into the EU8 group and the EU 15 group. As can be seen from
the graph, both the treatment and control group have an increase in output following the EU
expansion in the range from the year 2003 to the year 2004. While both groups have a similar
increase in production in absolute terms- around 600 mio. Euros, the EU8 group has an average
of 19% increase while the EU15 group has a 3% increase over the same time span.
Figure 7
20000.00
. 18000.00
16000.00
-·-··-----·--·-'
' .$ ~ ''~"'-~"' "......... • .,
. -~-- --~ -~-- -- ----
-·--I I ---·"·- .. , ...•... I -- ---------- I -
I I I ·--------- --------------1 I
14000.00 -1----------T--
1
I I ·---.---------------------
12000.00
10000.00
8000.00
6000.00
4000.00
2000.00
0.00
I I I I -,
~----
·---- ·---~--~---- -----
_J _____ -I I
'
1 I I
-----}--I I I
--------I-I I I
1 I I
I I I
·----1 I I
-"-- --- ---- I I ·----f----------·- ----------------------
1 I
: I I
J..------------------!---------------J.-..--------~----- ---- -------------- -·· I
I I -: --~-- - ----~-- ------:--- ----~----~- -- -
I : : ---t---·---~--~. -----.-•--,----·---- ------
7007 /00.~ ./004 ./001)
--EU8
- ··· Ell 1 S
t__ __________________________ ------------·--·---·· ·-. --------
When looking at the variation of data in the different variables box-plots provide a useful
way of graphically illustrating the relative concentration of data. We can here see two box plots
that for two important variables, agricultural production proportion of organic farming (box-plots
for all variables can be found in the Appendix). When looking at the box-plot for agricultural
production we can sec that the interquartile range from q 1-q2 is a lot less pronounced than the
other two interquartile ranges. Despite there being outliers for both agricultural production and
organic crop area proportion, the data is reasonably varied. When looking at organic crop area
proportion specifically we sec the opposite, namely that the interquartile range from q3-q4 is
truncated, even so looking at the ranges and dataset as a whole there is no reason to reject the
results or data because of lack of data dispersion. One variable did not show a lot of variation,
namely average fann size which was mostly because there are low numbers associated with the
11 Western European countries and high numbers associated with the EU8 population.
Figures 8 and 9
Oganic Crop Area Proportion AgrirulturaJ Production Statistics
N 18 18 18 18 N 18 18 18 18
"''" 0 0 1 2 l.lln 5110 5110 677 751
"'•• 27 27 47 60 "'" 8.3,757 a2,611 64.,4.77 63,567
02 11 12 12 u 02 6,5111 6,717 7,004 6,584
!Aoon 11 13 u 17 Ueon 1t,S71 14.518 15,131 U,U7
eo I'!<XlOO
or,. 0 rn> Drra or,.
"" Op.,. oocoo
0 p
' ' g 0 o •• a Dp.,. d n ~ «XXX> I 40 c I
A I
' 0 ..
-~~ ~ 9 n
a s = § ll l'l
~ 20
0 8 ~ 0 =
V:..... Findings
The first regression run was Ln (Y;r) = ,\0 + ,\1 In amount of fallow land;r + ,\2 ln proportion of organic r· .. "'
farming;r + ,\3 1n average farm size;r +' ,\4 In GOP per capita;r + ,\5 EU variable;r + a In L;r B In K;r +6; + e;r
testing the significance magnitude of each independent variable on agricultural production- holding
labour and capital constant:
Table 2: Regression Results
Variable Degrees of " Parameter Estimate Standard Error T Value P Value Freedom
Intercept 1 7.221181 2.3374 3.09 0.0034 ""
lrrigable Area (ha) 1 0.019842 0.0146 1.36 0.1811
Labour (1000 AWU) 1 0.270108 0.1522 1.77 0.0824
Proportion of Organic farming(%) 1 0.036507 0.0129 2.82 0.007
Fallow land ;
(ha) 1 -0.03991 0.0165 -2.42 0.0193
Av. Farm
Size (ha), 1 0.389279 0.1054 3.69 0.0006 .
GDP Per
Capita (€) 1 0.00226 0.1703 0.01 0.9895
EU Dummy 1 0.090498 0.0225 4.02 0.0002
As can be seen from the results of the first regression, most of the independent variables were
statistically significant. The R squared score was very high, above 99%, but there are two
reasons to heed caution in assessing this model based on that statistic. The first of which is that it
is a time series analysis, and as such it is not uncommon to have very high r squared scores. The
second is that the output included 17 cross sections which were all siatistically significant at the
0.01 level. These cross sections capture country specific effects, such as weather, which is not
included in any of the independent variables.
Looking specifically at the inputs of Labour and Capital we can see that the parameter
estimates have a positive sign like we would expect, but especially Capital measured by Irrigable
Area docs not seem to be particularly statistically significant. Because we would expect the
coefficient to be positive, the p-value should be halved and the variable is thus just significant at
the 10% level. However, since Capital is such an important factor in production it is still
surprising to see such a relatively high p-value. This raises some concern of the accuracy of the
proxy of Irrigablc Area to represent Capital as a whole. It also causes some concern of
multicollinearity as this could affect the T-stats and thus also the p-values.
The next variable is proportion of organic farming. As can be seen readily from the
summary statistics table, the proportion of farming meeting the criteria for organic farming
increased in both EU8 and EU15, though it doubled in the EU8 while only marginally increasing
in the EU 15 group. This variable is statistically significant at the 1% level and interestingly
enough has a positive parameter estimate, indicating that higher levels of percentage of farming
being organic, holding inputs constant, will result in higher levels of agricultural production.
This is very curious, as we would expect this to be a negative coefficient as outlined in the
theoretical section. Possible explanation accounting for this statistically significant reverse result
could be because the organic farming output market is still relatively new and the incentives
made by the CAP is necessary for farmers to enter the market since there are significant barriers
to entry and risk associated with price variation in the agricultural output market. However, from
these results, it seems clear that that the higher levels of organic farming resulting from entering
the EU actually positively affects agricultural production holding inputs constant.
Looking at the fallow land policy variable, the amount of fallow land decreased in both
the EU8 area as well as the EU15 area, but it decreased much more drastically in the EU8 area as
can be summarized from the summary statistics table. The theoretical section outlined why an
increase in fallow land would lower output produced as well as agricultural production
(measured in mio. €) if the elasticity of demand is greater than I. Looking at the regression
output the variable is statistically significant at the 5% level and has a negative coefficient. This
would mean lower levels of fallow land positively affects agricultural production and by
extension that the elasticity of demand for food is perhaps greater than one. Another possible
explanation as to why it has a negative coefficient is that there arc still significant expenses
associated with having the land such as property taxes, maintenances expenses etc. which arc not
outweighed by the subsidy received by farmers for leaving the land fallow. Regardless of the
relative importance of variables explaining the reasoning behind the negative coefficient, since
entering the EU countries have reduced the amount of fallow land which increased the (short
run) productivity of the agricultural sector of the EU.
The next variable in the regression is the Average Farm Size variable. Upon entering the
EUthe EU8 group's average farm size increased by 3%, while the average farm size for the
EU15 group stayed constant. In theory, we would expect to see an increase in productivity from
increased returns to scale since big farms have lower average fixed costs as well as lower
average variable costs from bulk buying and specialization. The results seem to substantiate the
theoretical work as the variable is statistically significant at the I% level and has a pretty high
magnitude in the positive coefficient. The relative large magnitude reflects in partial the low
average farm size in ha for most EU15 countries (the average being 9), so in increase of one
hectare on average is relatively large increase.
The GOP per capita variable was anticipated to have a negative coefficient since higher
levels of GOP per capita arc associated with more stringent regulations on pesticides, water use,
fertilizers, plant growth regulators etc. which would reduce productivity. However, it is also
possible these same regulations would increase productivity if pesticides are in fact overused and
regulations would bring the use of pesticides closer to the optimum use point. The results
however indicate that neither of these theories is supported by the data available. The p-value is
99% indicating that there is a 99% chance that the parameter estimate is caused by random
variation in the data and not a relationship between GOP per capita and agricultural production.
The last variable examined is the EU dummy variable controlling for whether or not the
country is a member of the EU. What this variable captures which is not expressed in any of the
other independent variables is the amount of subsidies received by farmers in terms of both
direct subsidies, price stabilization subsidies as well as additional mobility of labour etc. This
variable is statistically significant at the 1% level and has a positive coefficient indicating that
being a member of the EU increases agricultural productivity as production increases while
labour and capital are constant.
Included in this regression was a test for constant returns to scale, which demonstrated
that there is decreasing returns to scale since the output elasticities do not add up to one. This
means that a 20% increase in inputs will result in less than a 20% increase agricultural
production, ceteris paribus.
_ '. -. The lack of statistical significance for_especially capital does seem to be problematic
when looking at the reliability of the results. There is also some concern for multicollinearity,
and a test measuring the Variance Inflation Factor for each variable shows how all variables
except capital has a VIF under S.except for capital. Normally VIF scores above 5 should cause
some concern with regards to the reliability oft-values while VIF scores above 10 indicate
serious danger of unreliable t-values. In an attempt to mitigate some of these potential problems
the capital and labour variables were substituted with a variable measuring the capital/labour
ratio.
Ln (Yit) = A0 + A1 In amount of fallow landit + A1 1n proportion of organic farming,, + t\ 3 1n average farm size,,
+ A4ln GOP per capita;1 + A5 EU variable it+ (a In L/61n K) +O; + e;1
Table 3: Regression Results
Variable Degrees of Parameter Standard Error T Value P Value Freedom Estimate
Intercept 1 10.99063 1.4145 7.77 <.0001
Capital/Labour Ratio 1 0.031984 0.0137 2.34 0.0237
Proportion of Organic farming
(%) 1 0.043218 0.0129 3.35 0.0016
Fallow land (ha) 1 -0.04521 0.0168 -2.7 0.0096
Av; Farm Size
(ha) 1 0.3091 0.1004 3.08 0.0034
'·h ''
GOP Per Capita (€) 1 -0.19099 0.1443 -1.32 0.1918
' '
EU Dummy 1 0.092099 0.0232 3.97 0.0002
Overall these results verify the findings from the first regression, all the variables which
were found statistically significant were also statistically significant in this model with the same
signs for the parameter estimates. Similarly, GOP per capita is also statistically insignificant
even though the p-value is significantly lower. In this model all variables had VIF scores below 5
indicating the significance levels of the previous model were not too disturbed by
multicollinearity. Of course, it should be noted that the capital-labour ratio is an imperfect
substitute in this context since this ratio can stay constant while both inputs increase as long as
they increase at the same rate. Thus, when looking at these variables it is not certain that the
levels of input are held constant, but since both models give very similar results it seems
reasonable to conclude both that the multicollinearity in the first model is not an issue and that
the capital-labour ratio is a decent substitute for the variables individually. Additionally, both
models were tested for hctcroskedasticity and the null hypothesis ofhomoskedasticity was
accepted.
VI - Conclusion
In conclusion it can be determined that the population of Central and Eastern European countries
joining the EU in 2004 benefitted not only in terms of market access and political security, but
also in terms a higher level of agricultural productivity combined with a higher level of
subsidies. There are three major components to the indirect effects of the Common Agricultural
Policy which all caused this: the fallow land policy, the incentive to do organic farming, and the
increased farm size. When discussing the benefit of the fallow land policy it should be noted this
would reduce agricultural productivity for any country joining the EU which doesn't already
have fallow land subsidy greater than that of the CAP. With regards to organic farming, higher
proportions of organic farming cause higher levels of agricultural productivity because the prices
in the organic output market arc stillso high that the higher revenues received by suppliers still
outweighs the reduction in absolute crop, yield. One of the potential reasons why the transition
into the organic farming market has not equilibrated in the short run is because of relatively high
barriers to entry and price vagaries in the organic output market. The third variable is the Euro
per ha policy which favours larger farmers who are more efficient than small scale farmers
because of lower average fixed costs as well as lower average variable costs. Outside of these ' '
direct effects, being a member of the EU also increases agricultural production while controlling
for labour and capital inputs, primarily because of the direct subsidies promoting increasing
returns to scale. In terms ofGDP per capita this variable was not discovered to have a great
impact, but as indicated earlier the proposed relationship of the Environmental Kuznets Curve
would be easier to spot over a long term time series dataset. It should also be noted here that
while the CAP does seem to increase productivity through the mechanisms elaborated. this
comes at a hefty price tag of over € 40 bn. Whether the CAP itself is justified is a question which
lies outside the realm of this study, but some of the findings here may give clues to how such a
question can be answered in looking at some of the specific components of the CAP.
Future investigation of the relationship between the CAP and agricultural productivity
could with benefit include organic beef and pork production in addition to just organic crop
production. Moreover, a longer time range than the four years used in this study could establish
some of the patterns exhibited here more clearly. Furthermore, a future study would also be able
to include Bulgaria, Romania, and perhaps other countries joining the EU as separate treatment
groups. Perhaps this study would then also be able to effectively support or reject the validity of
the application ofthe Environmental Kuznets Curve to pesticide usc and environmental
regulations in the agricultural sector. Further studies could also include climatic variables such as
temperature, precipitation, wind etc. However, methodological obstacles such as the non-linear
relationship of many climate variables as well as the geographic size and climatic distributions in
countries remain.
Finally, some cautionary notes should be given regarding the findings in this study. First
of all the data used was, as described in section IV, interpolated. This always involves dangers
for the validity of the results due to data inaccuracy. The dataset was also rather small in terms of
years covered, which means overarching patterns may not be clearly deciphered in this study.
Moreover, the capital proxy was imperfect; perhaps a weighted capital variable combining all
farm instruments with irrigation materials could depict the capital used more accurately.
I pledge to have neither given nor received unauthorized assistance during the completion of this
work:
Flemming Schneider Rhode
N, , , , Min ...
Max
02
Mean
p r 0
d u 40XX) c t i 0 n
0
N
Min
Max
02 Mean
4(XX)-
F a I I 0 20JO-w L a n d
0
Appendix.
Agricultural Production Statistics
0 Fra
18
580
63,757
6,581
14,571
0 Fra
18
589
62,611
6,717
14,518
0 Fra
0 Ita
18
677
64,477
7,004
15,131
Fallow Land Summary Statistics
18
1
3,195
207
552
0 Spa
0 Pol
18
1
3,353
213
531
o Spa
18
1
3,273
191
487
0 Fra
D Spa
18
751
63,567
6,584
14,447
1B
2
3,319
205
4-BO
0 Pol
------~---------~:----------~---------------~': ________ _ I I
2C04
Organic Crop Area Proportion
N 1B 1B 1B Min 0 0 1 Max 27 27 4.7 02 11 12 12 Mean 11 13 14.
oo-
60 0 r g a 0 Por n i 40-c A r e a
20
-=-
a - --------------------------I--------------------------------I--------------------------------1 I I
N
Min
Max
02
Mean
I ~-r r i g a b 2000CXX)-
I e A r e a 1CXXXXJO
2003
lrrigable Area Statistics
1B 18 18
124. 74.0 970
3,979,4.80 3,977,210 3,974.,94.0
231,370 219,000 196,033
820,005 815,018 810,099
0 Ito 0 Ito 0 Ito 0 Spa 0 Spa 0 Spa
0 Fro 0 Fro 0 Fro
-- -r- 0 Gre
0 Por
-
I
2005
0 Ito 0 Spa
0 Fro
0 Gre
1B
2
60
14.
17
18
790
3,972,670
173,570
805,181
0 _____________ ;; ___________ ; __________ Q ___ Q ______ _ I I I I
2004 2005
Yea-
Labor Force Statistics N 18 18 18 18
Min 12~ 7~0 970 790
Max 3,979.~80 3,977,210 3,974,9~0 3,972,670
02 231,370 219,000 196,033 173,570
Mean 820,005 815,018 810,099 805,181
4aXXXX) D Ito D Ito 0 Ito D Ito D Spa D Spa D Spa D Spa
I m r D Fro D Fro 0 Fro D Fro r i g a b 20CXXXX)
I e
D Gre 0 Gre A -.- -.-r e a 1<XXXXJO
0- ··-··-·-···;;·····--·····--···;·-···--····--g···---·············8···--··-1 I I I
2(X)2 2(X)3 2CX)4 2(X)5
Year
Farm Size Statistics
N 18 18 18 18
Min 1 1
Max 375 353 359 390
02 6 6 6 6
Mean 30 29 29 31
400-D Ut
D Lit D Lit D Lit
3X>
f a r m 2CX) s
z e
100
0 -------------------------~------------------------------~-------------------------------~~-~---------------------------~-~~---------------------1 I I I
2(X)2 2CXX3 2004 20)5